Climate Risk, Risk Management, and the Allocation of U.S. Cropland: A Land Demand Based Approach

U.S. farmers have reallocated land across crops on an extraordinary scale over the past thirty years, and how much of this reflects climate adaptation matters for the welfare cost of warming. A farmer who switches from cotton to soybean may be chasing higher expected returns or fleeing rising downside risk. The two motives carry very different welfare and policy implications but reduced-form acreage-share methods cannot tell them apart. We build a structural land-use framework in which risk-averse farmers choose among crops by maximizing CARA expected utility over believed non-Gaussian per-acre profit distributions. The believed distributions are constructed from a thirty-year rolling weather belief and a Just–Pope yield model. The acreageshare system estimates a behaviorally disciplined distribution of farmer absolute risk aversion from observed land shares, rather than calibrating it from surveys. Federal crop insurance enters explicitly through its indemnity-and-premium structure. The framework lets us decompose climate-driven between-crop reallocation into a productivity channel and a risk channel, evaluate the welfare value of letting farmers re-optimize across crops, and assess subsidized federal crop insurance as a climate-adaptation policy. We apply the model to U.S. county-level land-allocation data 1987–2022 and project a CMIP6 SSP2-4.5 climate forward to mid-century. Two findings emerge. Between-crop reallocation absorbs most of the welfare cost of climate change the model assigns. About one-third of the projected reallocation reflects substitution away from worsening downside risk rather than toward higher expected returns — a channel reduced-form acreage methods cannot identify — and the same one-third risk share appears retrospectively in the actually-observed 1990–2020 reallocation.